BE HEARD

The description states that BE HEARD is an Android accessibility and productivity prototype. It supports two input modes:

• Voice Mode: Uses Android speech recognition to create an editable transcript. Approved content is sent through a protected gateway to GPT-5.6 for structured output.
• Sign Mode: A constrained personal fingerspelling prototype using MediaPipe Tasks Vision to extract hand landmarks. It supports seven static configurations (A, C, I, M, O, T, U). Calibration data remains in app-private storage and can be audited or exported.

Both input modes feed into a shared GPT-5.6 language pipeline for interpretation and output generation. The system allows users to review, edit, and confirm results before saving or sharing.

Inference The product is not a finished commercial offering but a prototype built for a hackathon.

The description states that BE HEARD explores a more flexible approach to digital communication, where users can express intentions through voice or fingerspelling and remain in control of the final output. It positions itself as an accessibility tool that prioritizes user agency and editable results.

Inference The positioning is centered on accessibility and user empowerment. It does not claim to be a full replacement for existing communication tools or a commercial product.

The description does not explicitly identify target customers or personas. However, it implies that BE HEARD is intended for users who benefit from alternative input methods, such as those with hearing impairments or other accessibility needs.

Inference The ICP likely includes individuals who use sign language or require assistive communication tools. No explicit segmentation beyond this is provided.

No business model or pricing information is provided in the description. The project is described as a prototype, not a commercial product.

Inference There is no evidence of monetization or pricing strategy at this stage.

The system uses a hybrid architecture combining local perception (MediaPipe Tasks Vision, Android speech recognition) and cloud-based language understanding (GPT-5.6). It includes:

• Local data storage with SQLCipher encryption.
• Cloudflare Worker as a protected gateway for OpenAI API access.
• React Native, Expo, Kotlin, TypeScript stack.
• Privacy Preview feature that hides raw camera images.
• Explicit consent for cloud processing.

Inference The architecture prioritizes privacy and user control. It appears designed for a secure, localized-first approach to accessibility input.

There is no evidence of traction, revenue, or customer adoption beyond the prototype phase. The project was built during a hackathon and submitted as a competition entry.

Inference No data on usage, retention, or monetization is available. The product is in early-stage development.

The description does not mention competitors or the broader market landscape. It focuses on the prototype’s unique features and design decisions rather than positioning within an existing ecosystem.

Inference No competitive analysis is provided. The product does not appear to be part of a known market category or platform.

• Prototype-only: No evidence of commercial viability, traction, or monetization.
• Limited scope: Sign Mode supports only seven static configurations and does not claim full LIS coverage.
• No production-ready features: The gateway is described as temporary, not production-grade.
• Unproven user adoption: No data on how many users would adopt this tool or its effectiveness in real-world use.

Inference This is an experimental prototype, not a product ready for market.

1. What is the intended user base beyond the prototype?
2. Are there plans to expand Sign Mode beyond the current seven configurations?
3. How will the system handle multi-user or real-world usage scenarios?
4. Is there any plan to monetize or scale this product beyond the hackathon?
5. What are the technical and legal implications of using GPT-5.6 in a privacy-sensitive context?